Image Quality Subsystem for Search Relevance Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital search systems fail to effectively differentiate between high and low quality images in search results, impacting user perception and relevance scoring.
Innovation Solution
An image quality subsystem computes query-specific quality scores based on initial quality scores and a transformation factor, adjusting relevance scores to prioritize high-quality images in search results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If text-based relevance scoring is used to select images for search results, then image relevance to query is improved, but image visual quality differentiation is worsened
Solution Approach 1:
The patent segments the image assessment process into two independent components: text-based relevance scoring and image quality scoring. The relevance score measures how well the image content matches the query text, while the quality score separately evaluates visual attributes such as sharpness, exposure, and color accuracy. This segmentation allows both relevance and quality to be measured and optimized independently, resolving the contradiction between relevance measurement precision and quality assessment capability.
Solution Approach 2:
The patent introduces an intermediary image quality score that acts as a mediator between the relevance scoring system and the final search result selection. This quality score is computed independently using image processing techniques and then combined with the relevance score through a weighted formula. The intermediary quality metric enables the system to differentiate image visual quality without compromising the text-based relevance measurement, thus resolving the technical contradiction.
2Quantity of substance
If all relevant images are included in search results regardless of quality, then search result completeness is improved, but user satisfaction is worsened
Solution Approach 1:
The patent changes the parameter used for selecting search results from solely relevance-based scoring to a composite scoring system that incorporates both relevance scores and quality scores. By adjusting the weighting parameters in the composite score formula, the system can control the balance between result completeness and quality. This parameter change enables the system to maintain a sufficient quantity of search results while improving user satisfaction through enhanced visual quality, as high-quality images are prioritized in the final ranking.
3Manufacturing precision
If image quality scoring is added to the search system, then image visual quality is improved, but system complexity is worsened
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing image quality scores in an image quality database before the search query is processed. During search operations, the system retrieves pre-computed quality scores from the database rather than performing real-time image analysis. This preliminary computation separates the complex image quality assessment from the time-critical search retrieval process, reducing system complexity during query execution while maintaining high image quality measurement capability.
Solution Approach 2:
The patent creates a simplified copy or representation of image quality characteristics through numerical quality scores stored in a database. Instead of performing complex real-time image analysis during search operations, the system uses pre-computed quality metrics that capture the essential visual quality attributes. This copying approach reduces computational complexity during search while preserving the ability to differentiate and prioritize high-quality images in search results.
Data Source
AI summary
An image quality subsystem computes quality scores for images that represent a measure of visual quality of the images. Initial quality scores and query specific quality scores can be computed for the images based on image feature values for the images and a transformation factor that represents a measure of importance of image quality for computing relevance scores for images. The initial quality scores are query independent quality scores that are computed for the images and can be used as a factor for computing relevance scores for the image relative to any query. Query specific quality scores are computed for images that are identified as relevant for a particular query based on the initial quality scores and a query specific transformation factor for the particular query. Adjusted relevance scores for the images can be computed based on the initial quality scores or the query specific quality scores.